Page 133 - Haematologica June

MYD88L265P mutation detection by ddPCR on BM, PB and ctDNA
able screening of low infiltrated WM specimens, such as unsorted BM or PB (routinely used in clinical practice);
3. both BM and PB are informative in treatment-naïve patients when tested for MYD88L265P by ddPCR; however, in relapsed patients (mainly exposed to rituximab), PB has a 1 log lower median mutated:WT ratio compared to BM, likely related to the efficacy of the anti-CD20 antibody in clearing the circulating tumor cells: therefore, BM analysis should be preferred for these patients;
4. the MYD88L265P ddPCR assay can be reliably used for MRD purposes, being as informative as the classical, stan- dardized but less applicable, IGH-based MRD assay, not yet validated in WM;
5. the MYD88L265P ddPCR assay targeting ctDNA is very promising for identifying the mutation in less invasively collected tissues, such as plasma or urine, and in samples from pre-treated patients.
The here described MYD88L265P ddPCR assay showed an overall mutation detection rate on baseline unselected mononuclear cells samples of 95.3% in BM and 71.2% in PB. These data might help to solve the still open debate on whether to use sorted versus unsorted BM mononu- clear cells to assess the MYD88L265P mutation36 and to extend the implementation of the ddPCR mutation assay to general diagnostic laboratories that do not routinely perform cell selection. At present, only 40% of MYD88L265P cases can be detected by ASqPCR using unse- lected PB cells, with an overall sensitivity of 39.5% and specificity of 100%.37
In our study we observed that the ddPCR assay on uns- elected cells greatly improved the rate of MYD88L265P detection compared to that achieved by ASqPCR. In fact, an analysis of 74 paired BM/PB baseline samples showed an overall detection rate of 93% (69/74) in BM and 72% (53/74) in PB. Among 69 patients with detectable muta- tions in BM samples, the sensitivity for MYD88L265P muta- tion detection by ddPCR on paired PB samples was 77% (53/69). The highly sensitive results of MYD88L265P ddPCR assay on unsorted samples makes this assay ideal for diagnostic use in clinical routine, avoiding the costs and technical requirements of cell sorting. In addition, our data confirmed, as also reported by Xu et al.,37 that PB samples are suboptimal for mutational screening in previ- ously treated patients, evidence confirmed by the high rate of false-negative results. The sensitivity of our assay in PB samples dropped from 85% in treatment-naïve patients to 47% in relapsed patients. Thus, a BM sample is essential to accurately identify MYD88L265P WT status in pre-treated patients (strongly suggested, for example, before starting an expensive therapy with ibrutinib).
This study shows that, besides its potential diagnostic role, MYD88L265P can effectively and easily be used for MRD monitoring in WM, achieving similar results to the less- applicable IGH-based MRD assay. In fact, monitoring allele levels can also provide insights into treatment effectiveness in a disease whose therapeutic scenario is rapidly changing, with many new and highly effective drugs.13 Of note, a report claimed the achievement of the first “molecular remission” in WM treated with carfilzomib.38 However, the deepness of molecular response matters, as does the sensi- tivity of the assay used for MYD88L265P detection, which was quite modest in that case (1.00x10-3)6 compared to the newly developed flow cytometry assays39 and with the ddPCR strategy described in this manuscript.
We observed that the agreement between ddPCR and
ASqPCR results was weaker in follow-up samples than in baseline ones. Further evaluations on large, prospective series of patients are needed to assess whether ddPCR could be useful for the identification of false-positive ASqPCR results, as was recently demonstrated in a next- generation sequencing/real-time quantitative PCR com- parison performed in acute lymphoblastic leukemia.40 Moreover, a methodological validation against IGH- based MRD detection and multiparametric flow cytome- try, as well as correlations with clinical impact, are eager- ly awaited in this setting and are currently ongoing in series of external samples.
The most innovative aspect of our study is the impact that ctDNA might have on MYD88L265P mutation detec- tion and MRD monitoring. We showed that ctDNA mir- rors the BM mutational burden much better than gDNA from PB at baseline. Moreover, our data suggest that ctDNA might be able to reflect the dynamic changes in tumor burden in response to treatment, with higher sen- sitivity than PB, which is particularly promising in pre- treated cases (Figures 3 and 4). Similar promising data have been obtained with ctDNA extracted from exo- somes from plasma and urine in a pivotal series of patients (D Drandi et al., 2018, unpublished data), although further investigations, as well as technical opti- mization, are needed. Nevertheless, these data add to the previously published experiences on the promising role of ctDNA analysis in lymphoproliferative disorders other than WM.41–50 Indeed, ctDNA represents an alternative, less invasive and “patient-friendly” tissue source for mutational analysis, and eventually MRD, which is espe- cially attractive for screening and monitoring asympto- matic patients, such as those with MGUS. (MYD88L265P ddPCR screening on a large group of IgM-MGUS patients is currently ongoing). However, plasma collection needs particular care since blood collection tubes and drawing procedures, as well known, can affect the stability of ctDNA.33,34 Pre-analytical standardized procedures are a prerequisite for clinical application and for consolidation of the promising potential of ctDNA analysis.
Finally, we described the impact of different therapies on MRD clearance in WM (i.e. chlorambucil, rituximab monotherapy), RCD (rituximab, cyclophosphamide, dexamethasone), FCR (fludarabine, cyclophosphamide, rituximab) and bendamustine-containing regimens (Online Supplementary Appendix). We acknowledge that the limitations of the present study include the compos- ite and retrospective nature of the series of patients characterized, moreover, by little outcome information. One of the main goals of this manuscript was the tech- nical description of the MYD88L265P ddPCR assay on unsorted cells. The illustration of the broad range of possible applications, including those on liquid biopsy, makes this approach potentially practical for implemen- tation in routine diagnostic laboratories. We are aware that this assay represents exploratory research, whose real advantages and predictive values need to be further validated in the context of prospective clinical trials. For this purpose, we are currently investigating MRD assessment by ddPCR on gDNA and ctDNA in the con- text of a phase II prospective study of relapsed WM patients, sponsored by the Fondazione Italiana Linfomi (EudraCT number: 2013-005129-22). Moreover, further efforts to establish uniform guidelines regarding stan- dardization procedures for sample collection and
haematologica | 2018; 103(6)
1035